**The Transition to Adulthood in Europe at the Intersection of Gender and Parental Socio-Economic Status**
Authors: Valeria Ferraretto and Agnese Vitali, University of Trento (IT)

The .zip folder contains:
- the do-file used to conduct the analyses (both in .do and in .txt, named "Dofile_complete_Aug24")
- a random sample drawn by the original dataset that allows to replicate results (both in .dta and in .csv, named "ESS_sample").

Data has been analysed with STATA 18 (StataNow/SE 18.5). 
Previous versions of the software can be used as well.

For any inquiries, please feel free to contact the corresponding author: valeria.ferraretto@unitn.it

*-------------------------------------------------------------------------------------------------------------------------------*
DATA AVAILABILITY
This paper uses individual-level microdata collected from the European Social Survey (ESS).
ESS data can be freely downloaded, upon login, at the following address:

https://ess.sikt.no/en/?tab=overview

The following rounds have been used:
ESS Round 3 - 2006
ESS Round 9 - 2018

Documentation is available for each of the rounds, also without logging in.
Data can be downloaded in csv, spss, or stata format.

A sample dataset from Round 9 and Round 3 (9000 cases) has been provided, both in .dta and in .csv format, for illustrative purposes.
The variables used in the analysis are listed below.

The timing of events of the transition to adulthood can be retrieved from:
- lvpntyr (first left the parental home)
- evpdemp pdempyr (year of first job, ever had a job)
- evlvptn lvptnyr (first coresidence with partner, ever coresided)
- fcldbrn bthcld (first child, ever had a child)

Control variables include:
- gndr (gender)
- edulvlfa edulvlma edulvlfb edulvlmb (parental education, varying depending on the round)
- emprf14 emprm14 (parental employment status)
- essround
- yrbrn (year of birth)
- agea (age at interview)
- cntry (country of residence)
- dweight pweight (design weight, population size weight)

*-------------------------------------------------------------------------------------------------------------------------------*
STEPS FOR REPLICATION

Download the 2 rounds from the ESS website and create a unique data file.

The do-file allows you to:

A. PREPARE the data
- for each event (first home-leaving event, first job, first coresidential union, first child), compute failure and duration indicators required for event-history analysis, and define criteria to select the sample. This will allow you to analyse survival times to each of the 4 events, using 4 comparable samples (selected using the same criteria).
- prepare the relevant covariates and control variables, common to all analyses (e.g., parental education, birth cohort)
 
B. ANALYSE the data
Each event has been analysed separately.
To this end, the full dataset prepared in step A should be (re)loaded when analysing a new event.
The analysis of each event includes:
- Transformation of the dataset in person-years format 
- Kaplan-Meier estimates (plot by cohorts and gender, Fig. 1; median ages by parental SES, gender, and country group, Table 3 and Table A1)
To produce Table 3 and A1, estimates for each event have been copied in a separate .xls file.
- Logit model by country group (M1) and AMEs (Fig. 2)
- Logit model by country group, with interactions (M2) and AMEs (Fig. 3)

In total, 16 models should be estimated (4 events x 4 country groups) from M1 and M2, respectively.
To produce Fig. 2 and Fig.3, make sure to run the models on all the events, by each country group, and store the results.
Each figure conveniently summarize results from 16 different models.